Research teams often hand us a report and ask which part is the story. Usually it is not the headline finding; it is something two tables further down. These are the questions we work through.
Is the question one people already ask?
Search for the question in plain words. If journalists, forums and search engines are already asking it, the answer has an audience. If nobody is asking, the story needs to explain why they should.
Does the data answer it, or only hint at it?
A sample of 200 respondents in one city cannot carry a national headline. We would rather publish a smaller claim that stands up than a larger one that gets corrected.
What changed?
Comparison over time is the simplest story structure there is. If the measurement has been taken before, the difference is the story. If it has not, say so: a first measurement is also news.
Who is affected differently?
Split the result by age, region, income, company size, whatever the data allows and the sample supports. The gap between two groups is often more interesting than the average.
Can it be shown in one chart?
If the finding needs three charts to make sense, it is three findings. Pick one.
Would you defend it on the phone?
Every story we publish has a named researcher who will take a journalist's call. If no one on the team is comfortable defending the claim, it is not ready.
The best research stories are not the biggest numbers. They are the ones a reader can repeat at dinner without getting them wrong.
After the checklist
Once a finding passes all six, we write it up in plain language, build the chart, and start the outreach. The whole process from dataset to first placement usually takes four to six weeks.